MSc Thesis Proposal

Model-Based Predictive Control
for Distributed Energy Resources


Mentors: Rudy Negenborn, Bart De Schutter, Igor Kluin (Qurrent)

Keywords: Model-based predictive control, distributed energy resources, renewable energy, household optimization

Description:
Distributed energy resources, such as small-scale power generators, electricity storage units, and responsive loads, can play a crucial role in combating climate change, increasing the amount of electricity generated from renewable sources, and enhancing energy saving. In the near future, households will be equipped with small-scale power generators, such as solar panels and wind turbines, and with energy storage devices, such as batteries and heat storage units. Households with such distributed energy resources can operate more independently of energy suppliers, and they can buy and sell power both among one another and to and from their supplier.

Neighborhood.
Fig. 1. A small local energy network.

Also, households will be equipped with devices that can monitor and switch on and off energy production and consumption. These devices will allow more efficient control of energy flows within households. A high-tech startup company that is developing technology for enabling this is Qurrent. Qurrent develops devices, software, and services that enable the creation of small local energy networks, such as shown in Fig. 1. The households in the network exchange energy to maximize the efficiency of the energy they produce.

Qurrent has developed the Qbox, the Qserver, and the Qmunity website. The Qbox, as shown in Fig. 2, is a small box that is included in each participating house or office building in the local energy network. The data from the Qboxes of several households can be collected at a central location. At this central location, it can de determined how the Qboxes should control the energy appliances of the households. The Qserver performs this control task. It therefore includes a control algorithm that determines how the Qboxes should manage energy flows. The Qmunity website, as shown in Fig. 3, shows the user energy performance information through easy to understand graphs.


Qbox. Qmunity.
Fig. 2. The Qbox.Fig. 3. The Qmunity website.

Assignment:
In this MSc project you will focus on the control algorithm for the Qboxes. You will work on developing a control strategy that can run on the Qserver and that can coordinate the energy consumption and production of the participants in the local energy network through the Qboxes. We will start by making a description of the control problem to be solved and a survey of appropriate control techniques. Then we will select one of the techniques and use this technique to design a controller for the Qserver. Finally, there is the opportunity to implement the resulting controller on a real-life setup.

This MSc project will partly be done at TU Delft and partly at Qurrent (located in Schiedam, 10 minutes by train from Delft).


If you are interested in selecting this project as your MSc project, please stop by at the office of Rudy Negenborn or Bart De Schutter, or send an e-mail to r.r.negenborn@tudelft.nl for more information.